AI Research / 2026

Explainable Legal Judgment Prediction for Indonesia

Undergraduate thesis research at LEXIN pairing LLM understanding with verifiable symbolic reasoning to predict Indonesian court verdicts with explanations that show their work.

Legal NLPNeuro-symbolic AILLM evaluationResearch design

At LEXIN (Center for Legal Informatics, Universitas Indonesia), I am researching explainable legal judgment prediction for Indonesian criminal decisions. The design pairs an LLM that extracts structured claims with verbatim citations from masked court decisions against a deterministic symbolic engine that adjudicates verdicts over hand-curated statutory element tables.

The work so far includes a systematic literature review with around 210 new paper captures across ACL, EMNLP, ICAIL, JURIX, and NLLP venues, an 18-paper core corpus read and citation-verified in full, a mapping of 20 Indonesian legal NLP works, and the Chapter 2 related-work synthesis plus Chapter 3 method positioning — establishing a composition-level research gap no prior system covers for Indonesian criminal decisions.